Assimilation of SMAP Soil Moisture Retrievals into the WRF-Noah Model Under a Framework of Strongly-Coupled Data Assimilation

Remotely-sensed soil moisture data has been incorporated into numerical weather models for improving weather forecasts. The most common way is via a weakly-coupled data assimilation framework, with which the soil moisture data assimilation is only seen via the land-surface model integration. A strongly-coupled data assimilation requires the estimation of cross-model background error covariance and the simultaneous correction of atmospheric and land surface model states during the analysis procedure. Due to uncertainties in numerical models in representing the land-atmospheric interfaces, strongly-coupled data assimilation becomes a challenging problem. This study examines the effectiveness of a strongly-coupled data assimilation through assimilating data from the Soil Moisture Active Passive (SMAP) on the weather forecasts and compares the results with a weakly-coupled data assimilation framework. In this study, we assimilate 9-km SMAP level-2 enhanced soil moisture data into the WRF-Noah model using a variational approach. Prior to data assimilation, the SMAP data were rescaled to the model space depending on the soil types. We conducted an open-loop experiment (OPL), and two SMAP data assimilation experiments, one that updates only surface soil moisture state (SDA) and the other that updates surface soil moisture and atmospheric temperature and specific humidity states (CDA) during the analysis procedure. Three-day forecasts of temperature and humidity profiles over the Great Plains are evaluated. In OPL, there is a bias of 1.76 K and -0.55 g/kg on average for temperature and humidity, respectively. SDA leads to a bias reduction of 7.8% and 20.5% for temperature and humidity, respectively, while CDA contributes an additional reduction of 2% and 4%. It is also found that SDA reduces the RMSE of temperature by 4.5% while CDA provides an additional RMSE reduction of 0.7%. However, the impact of SMAP data assimilation of specific humidity in terms of RMSE is marginal (less than 1% reduction).